JP2002010106A - Device and method for reducing noise and medium storing noise reduction program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a noise-reducing device that reduces the persistence of vision and non-clarity in contours according to the motion of an image signal and an image signal level for improving S/N ratio, a noise reduction method, and a medium storing a noise reduction program. SOLUTION: A frame cyclic type noise-reducing device 1002, non-cyclic type noise-reducing device 1003, and a controlling device 1004 are provided. The control means 1004 controls the frame cyclic type noise-reducing device 1002 and non-cyclic type noise-reducing device 1003 according to the motion detection signal that is obtained from the frame cyclic type noise-reducing device 1002 and the image signal level, and the noise is reduced by making the best of both characteristics, thus removing the persistence of vision in motion reducing the non-clarity in contours in a stationary state, and at the same time mainly reducing the noise of a signal where an image signal level is small.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video signal processing apparatus, and more particularly to a noise reduction apparatus for reducing noise of a video signal and improving S / N.

[0002]

2. Description of the Related Art Generally, a video signal is a signal repeated at a frame period, and the autocorrelation between frames is strong. However, since random noise of a video signal has low autocorrelation, noise can be reduced by calculating a video signal at a frame period and replacing pixels.

Conventionally, a noise reduction device that performs such an operation has been a frame recursive noise reduction device provided with means for controlling based on motion detection information (JP-A-6-22).
No. 5178).

FIG. 11 shows a conventional frame recursive noise reduction apparatus based on motion detection.

As shown in FIG. 11, a frame recursive noise reduction apparatus includes a two-dimensional LPF (low-pass filter) 1102 for improving the accuracy of motion detection, and an adder 110.
3, a motion determination circuit 1104, multipliers 1105 and 1106 for obtaining cyclic and non-cyclic video signals, an adder 1107 for adding video signals, and a frame memory 110.
8 is provided.

[0006] In the frame recursive noise reduction apparatus, the input video signal 1101 is subjected to noise removal by a two-dimensional LPF 1102, and the obtained video signal and a frame memory 1108 are removed.
Is output to the adder 1103 and a difference is generated.

The difference calculated by the adder 1103 is
It is input to the motion determination circuit 1104, and it is determined whether or not the component is a motion component between frames. The difference includes a noise component having no inter-frame autocorrelation and a difference of a signal component changed between frames. Generally, the smaller the smaller, the higher the probability of a noise component, and the larger the larger, the higher the probability of a motion component. Get higher. Therefore, the motion determination circuit 1104
Then, if the difference is larger than the threshold value, the motion component is determined. If a plurality of threshold values having different sizes are provided and compared with the difference, the magnitude of the motion can also be determined.

When the motion is determined to be large by the motion determining circuit 1104, the value of the output of the motion determining circuit (K:
Control is performed such that 0 <K <1) is reduced so as to approach 0, and when the motion is small, the control is increased so as to approach 1.

The output of the motion judging circuit output 1104 indicating the motion of the input video signal 1101 is
Adjust the multiplication coefficients (1-K) and (K) of 5,1106,
By adding the two by the adder 1107, when there is motion, the input video signal 1101 is strengthened to reduce the effect of the frame-delayed video signal, and when there is no motion, the input video signal 1101 is weakened and the frame is delayed. By replacing the video signal by increasing the influence of the video signal, it is possible to reduce the noise of the video signal while reducing the afterimage.

As a method for reducing the noise of the video signal in a non-recurring manner without using the frame memory as described above, a method of performing a one-dimensional filter or a two-dimensional filter process is known. As a one-dimensional filter, a filter that averages adjacent pixels is known. Further, as one method of noise reduction, a coring method is known in which a signal below a certain threshold value is set to zero. As a two-dimensional filter, there is a method of performing a low-pass filter process in a horizontal direction and a vertical direction.

A method has also been proposed in which noise can be reduced by adaptively selecting and averaging pixels (Japanese Patent Laid-Open No. 6-86104). this is,
The filter central pixel and its peripheral pixels are taken in, the difference between the peripheral pixel and the central pixel is calculated, and the difference is compared with a threshold value for determining whether or not there is a correlation with the central pixel. If the difference is smaller than the threshold value, there is a correlation. By calculating the average of only peripheral pixels having a correlation and replacing it with the center pixel, noise reduction can be performed while reducing contour blurring without averaging contours with low correlation.

FIG. 12 is a block diagram showing a circuit configuration of a conventional noise reduction device for reducing noise of a video signal on a non-recurring basis.

As shown in FIG. 12, a noise reduction device includes a pixel fetch circuit 1202 and a difference calculation circuit 1.
203 to 1206, a threshold value 1207, and a correlation determination circuit 1
It is composed of 208 to 1211, a correlated pixel adding circuit 1212, a pixel number counting circuit 1213, and a divider 1214, receives an input video signal 1201, extracts a central pixel and its peripheral pixels, and calculates an average value. The output video signal 1215 is calculated and replaced with the center pixel.

The operation of each block will be described below.

An input video signal 1201 is input to a pixel capturing circuit 1202, where the central pixel and its peripheral pixels are captured, and the central pixel and each peripheral pixel are output to difference calculating circuits 1203 to 1206, respectively.

The difference calculation circuits 1203 to 1206 calculate the difference between the center pixel and each peripheral pixel, and the respective results are input to the correlation determination circuits 1208 to 1211. In the correlation determination circuits 1208 to 1211, the difference calculation circuit 12
The result inputted from 03 to 1206 is compared with the threshold value 1207, and if smaller than the threshold value, it is determined that there is a correlation,
The determination result is input to the correlation pixel addition circuit 1212.

In the correlated pixel addition circuit 1212, the central pixel and each peripheral pixel are input from the pixel capturing circuit 1202,
Based on the results determined by the correlation determination circuits 1208 to 1211, only the central pixel and peripheral pixel values having a correlation with the central pixel are added. Pixel counting circuit 1213
Counts the number of pixels correlated with the central pixel based on the determination results by the correlation determining circuits 1208 to 1211 and counts the number of pixels combined with the central pixel.

In the divider 1214, the correlated pixel addition circuit 1
The result of addition by 212 is divided by the number of pixels correlated with the central pixel counted by the pixel number counting circuit 1213 and the number of pixels correlated with the central pixel, thereby obtaining an average value of correlated pixels. This average value is output as an output video signal 1215 and is replaced with the center pixel.

According to the above operation, the contour portion having a low correlation is excluded from the target for calculating the average value, the blur of the contour portion can be reduced, and only the correlated pixels on which random noise is superimposed are averaged. And reduce noise.

[0020]

However, in such a conventional frame recursive noise reduction apparatus, when there is motion between frames, the amount of repetition is reduced to reduce afterimages, so that noise reduction is sufficient. There was a problem that was not.

Further, in the non-recursive noise reduction device, when noise is reduced by a filter process for averaging pixels, contour blur occurs. As for coring, a signal of a contour portion that is equal to or smaller than the threshold value is replaced with zero, resulting in contour deterioration. The method of adaptively selecting a pixel and replacing it with an average pixel also has a problem in that signal degradation occurs in a contour portion below a threshold value.

SUMMARY OF THE INVENTION The present invention has been made to solve such a problem, and a frame recursive type noise reduction device and a non-recursive type filter noise reduction device are combined to degrade the contour portion of a small video signal level. In the case where there is no motion between frames and the video signal level is small, noise reduction is mainly performed by the frame cyclic noise reduction device, and when there is motion between frames, It is an object of the present invention to provide an optimal noise reduction device according to the motion and level of a video signal by mainly performing noise reduction using a non-recursive filter noise reduction device.

[0023]

According to the present invention, there is provided a noise reduction apparatus which receives a video signal, adjusts a cyclic amount of the delayed image based on an image obtained from the video signal and a delayed image, and inputs the video signal. Recursive noise reduction device that reduces noise and outputs the same, and a non-recursive noise reduction device that receives a video signal from the recursive noise reduction device, performs a filtering process using peripheral pixels, and reduces and outputs the noise of the input video signal. Type filter noise reduction device, and control means for controlling the recursive noise reduction device and the non-recursive filter noise reduction device, wherein the recursive noise reduction device is based on the image and the delayed image. Motion determining means for detecting a motion of a video and outputting a motion detection signal; video signal level detecting means for detecting and outputting a video signal level of the input video signal; Cyclic amount adjusting means for adjusting the cyclic amount of the delayed image based on the motion detection signal and the video signal level; and Video signal output means for reducing the noise of the input video signal and outputting the same, wherein the control means is configured to control the recursive noise reduction device and the non-cyclic noise based on the motion detection signal and the video signal level. It has a configuration characterized by controlling the type filter noise reduction device.

With this configuration, a recursive noise reduction device for reducing noise of an input video signal and outputting the same, and a non-recursive filter noise for inputting a video signal from the recursive noise reduction device and reducing the noise for output Since the control unit controls the reduction device based on the motion detection signal and the video signal level, there is no afterimage where there is motion in consideration of the motion of the video and the video signal level, and the level is small when stationary. This means that the signal can be mainly reduced in noise.

Further, the noise reduction apparatus of the present invention reduces the noise of the input video signal by adjusting the amount of circulation of the delayed image based on the image obtained from the video signal and the delayed image. And a non-recursive filter noise reduction which receives a video signal from the recursive noise reduction device, performs a filtering process using peripheral pixels to reduce noise of the input video signal, and outputs the resultant signal. A video signal before the noise reduction, a contour correction device for performing contour correction of the image and reducing and outputting noise of the input video signal, and a video signal output by the non-recursive filter noise reduction device. Video signal output means for generating and outputting an output video signal from the video signal output by the contour correction device, the cyclic noise reduction device, and the non-cyclic filter A noise reduction device and control means for controlling the contour correction device, wherein the cyclic noise reduction device detects a motion of a video based on the image and the delayed image and outputs a motion detection signal. Determination means; video signal level detection means for detecting and outputting a video signal level of the input video signal; and a circuit for adjusting the amount of circulation of the delayed image based on the motion detection signal and the video signal level. An amount adjusting unit, and a video signal output unit that circulates the delayed image according to the amount of circulation adjusted by the circulating amount adjusting unit, reduces noise of the input image signal, and outputs the image signal. Based on the motion detection signal and the video signal level, the recursive noise reduction device, the non-recursive filter noise reduction device, and the contour correction device It has a configuration in which characterized in that Gosuru.

With this configuration, a recursive noise reduction device for reducing the noise of an input video signal and outputting the same, and a non-recursive filter noise for inputting a video signal from the recursive noise reduction device and reducing the noise and outputting the reduced noise The control means controls the reduction device and the contour correction device for performing contour correction of the image based on the motion detection signal and the video signal level, so that the motion of the video and the video signal level are considered in consideration of the video motion and the video signal level. When there is no residual image, when the image is still, noise can be reduced by focusing on a low-level video signal, and the contour of the video can be sharpened without noise enhancement according to the motion and the level.

Also, the noise reduction apparatus of the present invention has an automatic gain control apparatus for inputting a video signal and automatically changing the gain while maintaining a predetermined output signal strength, and a video signal input from the automatic gain control apparatus. A recursive noise reduction device that adjusts the amount of repetition of a delayed image based on an image obtained from a video signal and a delayed image, reduces noise of an input video signal, and outputs the rejected image signal. A non-recursive filter noise reduction device that receives a signal, performs filter processing using peripheral pixels to reduce noise of an input video signal, and outputs the reduced noise, and the recursive noise reduction device and the non-recursive filter noise reduction device. Control means for controlling the automatic gain control device, wherein the automatic gain control device has gain amount output means for outputting a gain amount obtained by changing a gain of an input video signal, and A motion reduction unit that detects a motion of a video based on the image and the delayed image and outputs a motion detection signal, and adjusts a cyclic amount of the delayed image based on the motion detection signal. And a video signal output unit that circulates the delayed image according to the circulating amount adjusted by the circulating amount adjusting unit, reduces the noise of the input video signal, and outputs the image signal. The control means controls the recursive noise reduction device and the non-recursive filter noise reduction device based on the gain amount and the motion detection signal.

With this configuration, the video signal input by the automatic gain control device is automatically changed in gain while maintaining a predetermined output signal strength, and the video signal is input from the automatic gain control device to reduce noise. And a non-recursive filter noise reducer, which receives a video signal from the cyclic noise reducer and reduces the noise, and outputs the result, using a motion detection signal and an automatic gain control device to adjust the gain. Therefore, the control means controls the noise based on the above, so that the noise can be reduced in consideration of the motion of the video and the gain amount.

Further, the noise reduction apparatus of the present invention reduces the noise of the input video signal by adjusting the amount of circulation of the delayed image based on the image obtained from the video signal and the delayed image. And a non-recursive filter noise reduction which receives a video signal from the recursive noise reduction device, performs a filtering process using peripheral pixels to reduce noise of the input video signal, and outputs the resultant signal. Device, and control means for controlling the non-recursive filter noise reduction device, wherein the recursive noise reduction device detects motion of a video based on the image and the delayed image to generate a motion detection signal. A motion determining means for outputting;
A cyclic amount adjusting means for adjusting the cyclic amount of the delayed image based on the motion detection signal; and a noise of the input video signal by cyclically moving the delayed image according to the cyclic amount adjusted by the cyclic amount adjusting means. Video signal output means for reducing and outputting the signal, wherein the control means controls an adaptive range of a filter process performed by the non-recursive filter noise reduction device based on the motion detection signal. It has the following configuration.

With this configuration, the non-recursive filter noise that reduces the noise of the video signal input by the recursive noise reduction device and outputs it, and that receives the video signal from the recursive noise reduction device, reduces the noise, and outputs the reduced noise. Since the control unit controls the reduction device by adjusting the filter area based on the motion detection signal of the video signal, it is possible to perform effective noise reduction in consideration of the motion of the video.

Further, the noise reduction apparatus of the present invention reduces the noise of the input video signal by adjusting the amount of circulation of the delayed image based on the image obtained from the video signal and the delayed image. And a non-recursive filter noise reduction which receives a video signal from the recursive noise reduction device, performs a filtering process using peripheral pixels to reduce noise of the input video signal, and outputs the resultant signal. Device, and control means for controlling the non-recursive filter noise reduction device, wherein the recursive noise reduction device detects motion of a video based on the image and the delayed image to generate a motion detection signal. A motion determining means for outputting;
A cyclic amount adjusting means for adjusting the cyclic amount of the delayed image based on the motion detection signal; and a noise of the input video signal by cyclically moving the delayed image according to the cyclic amount adjusted by the cyclic amount adjusting means. Video signal output means for reducing and outputting the image signal, wherein the control means weights and processes an adaptive pixel of a filter process performed by the non-recursive filter noise reduction device based on the motion detection signal. It has a configuration characterized by this.

According to this configuration, the noise of the video signal input by the recursive noise reduction device is reduced and output, and the non-recursive filter noise that receives the video signal from the recursive noise reduction device, reduces the noise, and outputs the reduced noise. Since the control unit controls the reduction device by adjusting the weight given to the peripheral pixels by the motion detection signal of the video signal, it is possible to perform effective noise reduction in consideration of the motion of the video.

The medium on which the noise reduction method and the noise reduction program of the present invention are recorded includes a video signal inputting step of inputting a video signal, an image delaying step of holding and delaying and outputting an image, and A motion detection step of acquiring an image and detecting a motion of a video based on the image and the delayed image; a signal level detection step of detecting a signal level of the input video signal; and Determining a cyclic amount of the delayed image based on a video signal level; and determining a filter control parameter for performing a filtering process on a video signal based on the motion of the video and the video signal level. Determining a filter control parameter and controlling the amount of recursion of the delayed image based on the determined amount of recursion. A recursive noise reduction step of reducing and outputting noise of an input video signal, and performing a filtering process on the video signal output in the recursive noise reduction step by selecting peripheral pixels based on the filter control parameters. And a non-recursive filter noise reduction step of reducing and outputting noise.

Thus, the amount of control and the control parameters are determined based on the motion detection and the video signal level detection, and the cyclic noise reduction and the non-recursive filter noise reduction are performed. In consideration of the above, noise reduction can be performed without a residual image in a place where there is motion, and noise reduction can be performed with emphasis on a signal having a small level in a stationary state.

Also, a medium on which the noise reduction method and the noise reduction program of the present invention are recorded includes a video signal inputting step of inputting a video signal, an image delaying step of holding and delaying and outputting an image, A motion detecting step of obtaining an image from a signal and detecting a motion of a video based on the image and the delayed image; a signal level detecting step of detecting a signal level of the input video signal; And a cyclic amount determining step of determining a cyclic amount of the delayed image based on the video signal level, and a filter control parameter for performing a filtering process on a video signal based on the motion of the video and the video signal level. Determining a filter control parameter to be determined, and contouring the video signal based on the motion of the video and the video signal level. A contour correction parameter determining step of determining a contour correction parameter for performing a correct process; and a cyclic control for controlling a cyclic amount of the delayed image based on the determined cyclic amount to reduce noise of the input video signal and outputting the reduced image. Noise reduction step, and a non-recursive filter noise reduction that outputs the video signal output in the recursive noise reduction step by selecting peripheral pixels based on the filter control parameters and performing the filtering process to reduce noise. A contour correction signal generating step of performing the contour correction processing based on the contour correction parameter from the video signal output in the cyclic noise reduction step to generate a contour correction signal; and the non-cyclic filter noise reduction step An adding step of adding the video signal output in step (a) and the contour correction signal. It is characterized in.

Thus, based on the motion detection and the video signal level detection, the amount of recursion, the control parameter of the acyclic filter, and the control parameter of the contour correction are determined, and the recursive noise reduction and the acyclic filter noise reduction are performed. Since contour correction is performed, the motion of the video signal and the video signal level are taken into account,
Where there is motion, noise reduction is performed by a non-recursive filter without residual images, and when stationary, low-level signals are primarily noise-reduced by a cyclic filter, and motion blur is reduced and noise is reduced. The contour blur can be reduced.

Further, the medium on which the noise reduction method and the noise reduction program according to the present invention are recorded has a video signal input step of inputting a video signal, and a gain amount for maintaining the input video signal at a predetermined output signal strength. Calculating a gain amount to calculate, an automatic gain control step of amplifying the input video signal based on the gain amount, an image delaying step of holding and delaying and outputting an image, and an image from the amplified video signal. And a motion detecting step of detecting video motion based on the image and the delayed image; and determining a cyclic amount of the delayed image based on the gain amount and the video motion. Determining a filter control parameter for performing a filtering process on a video signal based on the gain amount and the motion of the video. A control parameter determining step, a cyclic noise reduction step of controlling the cyclic amount of the delayed image based on the determined cyclic amount, and reducing and outputting noise of the amplified video signal; A non-recursive filter noise reduction step of selecting a peripheral pixel based on the filter control parameter and outputting the video signal output in the noise reduction step to reduce the noise by performing the filtering process. I have.

Thus, the amount of control and the control parameters are determined based on the motion detection and the gain calculation, and the cyclic noise reduction and the non-recursive filter noise reduction are performed.
In order to improve the phenomenon that the low level video signal amplitude is increased by C (automatic gain control) and the S / N is deteriorated,
Low-level signals amplified by AGC can be focused on by mainly performing cyclic noise reduction in stationary areas and noise reduction mainly in non-recursive filters when moving. Becomes

Further, the medium on which the noise reduction method and the noise reduction program of the present invention are recorded has a video signal input step of inputting a video signal, an image delay step of holding and delaying and outputting an image, and A motion detecting step of acquiring an image from a signal and detecting a motion of a video based on the image and the delayed image; and a circulating amount determining determining a circulating amount of the delayed image based on the motion of the video. And a filter area control parameter determining step of determining a filter area for performing a filtering process on a video signal based on the motion of the video, and controlling a cyclic amount of the delayed image based on the determined cyclic amount. A cyclic noise reduction step of reducing the noise of the input video signal and outputting the same, and an image output in the cyclic noise reduction step. It is characterized in that a non-recursive filter noise reducing step of selecting peripheral pixels and outputs to reduce noise performs the filtering process on the basis of a signal to the filter region control parameter.

With this, the amount of recursion and the control parameters are determined based on the motion detection, and the recursive noise reduction and the nonrecursive filter noise reduction are performed. By performing filter processing by enlarging the range of pixels with noise, noise reduction with the effect of eliminating afterimages is performed, and when stationary, cyclic noise reduction is performed. In these two types of noise reduction, signals with low levels are emphasized Thus, noise can be reduced effectively.

A medium on which the noise reduction method and the noise reduction program of the present invention are recorded includes a video signal inputting step of inputting a video signal, an image delaying step of holding and delaying and outputting an image, and A motion detecting step of acquiring an image from a signal and detecting a motion of a video based on the image and the delayed image; and a circulating amount determining determining a circulating amount of the delayed image based on the motion of the video. Based on the motion of the video, based on the motion of the image, a filter weight control parameter determining step of determining the weight of the pixel to perform a filtering process on the video signal, the cyclic amount of the delayed image based on the determined cyclic amount A cyclic noise reduction step of controlling and reducing the noise of the input video signal, and outputting the noise. It is characterized in that a video signal and a non-recursive filter noise reducing step of weighting the surrounding pixels and outputs to reduce noise performs the filtering based on the filter weighting control parameters.

Thus, the amount of recursion and the control parameters are determined based on the motion detection, and the recursive noise reduction and the nonrecursive filter noise reduction are performed. By expanding the area of a certain pixel, that is, the area of the filter processing weight from the center pixel of the filter, noise reduction with the effect of eliminating afterimages is performed by a non-recursive filter where there is motion, and In some cases, a signal having a small level can be effectively reduced by the recursive filter.

[0043]

Embodiments of the present invention will be described below with reference to the accompanying drawings. (First Embodiment)

FIG. 1 shows a configuration diagram of a noise reduction device according to a first embodiment of the present invention.

As shown in FIG. 1, the noise reduction apparatus includes a frame recursive noise reduction apparatus 1002, a non-recursive filter noise reduction apparatus 1003, and a control unit 1004, and receives an input video signal 1001 and performs noise reduction. , And output the output video signal 1005.

Frame cyclic noise reduction apparatus 1002
Is for performing a noise reduction process on the input video signal 1001 based on the motion detection information and outputting the result. The non-recursive filter noise reduction device 1003 is connected in series with the frame recursive noise reduction device 1002, and adaptively selects and averages pixels output from the frame recursive noise reduction device 1002 by averaging. The noise is reduced and output as an output video signal 1005.

The control device 1004 receives the motion detection signal and the video signal level from the frame recursive noise reduction device 1002, and controls the frame recursive noise reduction device 1002 and the non-recursive filter noise reduction device 1003. The motion detection signal and video signal level are
Conventional Example It can be extracted from the signal input to the multiplier 1105 shown in FIG.

The motion adaptive frame recursive noise reduction apparatus 1002 detects video motion from the input video signal 1001 and increases the amount of recursion of the delayed image when the motion is small, and decreases the recursive amount when the motion is large. In other words, the control is performed so that the effect of noise reduction is greater as the camera is closer to the stationary state, and the afterimage is not generated as the movement increases. When the video signal level is low, the amount of circulation is large, and when the video signal level is high, the amount of circulation is small. From the above, for example, when arranged in descending order of the traveling amount,

Stationary and small level> stationary and large level> moving and small level> moving and large level ... Equation (1) As described above, control can be performed by the control device 1004.

The frame recursive noise reduction device 10
For a video signal for which noise reduction was not sufficient in 02, the control device 1004 controls the non-recursive filter noise reduction device 1003 to increase the noise reduction effect. That is, control is performed such that the effect of noise reduction increases as the moving image is closer to the moving image. When the video signal level is low, noise is conspicuous, and control is performed to increase the noise reduction effect. From the above, for example, if the noise reduction effect is arranged in descending order,

Moving and low level> moving and high level> stationary and low level> still and high level Expression (2) can be controlled in this way by the control device 1004 to obtain the output video signal 1005 Can be.

For example, when the non-recursive filter is coring, the threshold value is set larger in the left term of the equation (2) than in the stationary state, so that the effect of noise reduction can be more easily exerted. . In the case where the noise is reduced by averaging the pixels, the number of peripheral pixels to be averaged is increased. In particular, in the acyclic noise reduction device shown in the conventional example, n is increased to increase peripheral pixels to be averaged.

As described above, the noise reduction apparatus according to the first embodiment of the present invention is a frame recursive noise reduction apparatus 1002 that performs noise reduction on an input video signal 1001.
And a control device 1004 to which the motion detection signal and the video signal level obtained from the frame recursive noise reduction device 1002 are input, and the non-recursive filter noise reduction device 1003 controlled by the control device 1004 is provided with a frame recursive noise reduction device. Because it is connected in series to the device 1002,
In consideration of the motion of the video and the video signal level, there is no afterimage where there is motion, and when stationary, the noise can be reduced by focusing on a signal having a low level. Further, the control means 1004 is not limited to the noise reduction device described in the conventional example,
The present invention is applicable to various non-recursive filters, and provides an effect. (Second embodiment)

Next, FIG. 2 shows a configuration diagram of a noise reduction device according to a second embodiment.

As shown in FIG. 2, the noise reduction device includes a frame recursive noise reduction device 1002, a non-recursive filter noise reduction device 1003, control means 1004, a contour correction device 1006, and an adder 1007. 1001 is input, noise is reduced, and an output video signal 1005 is output.

Frame cyclic noise reduction apparatus 1002
Is for performing a noise reduction process on the input video signal 1001 based on the motion detection information and outputting the result. The non-recursive filter noise reduction device 1003 is connected in series with the frame recursive noise reduction device 1002, and adaptively selects and averages pixels output from the frame recursive noise reduction device 1002 by averaging. The noise is reduced and output.

The control device 1004 receives the motion detection signal and the video signal level from the frame recursive noise reduction device 1002 and controls the frame recursive noise reduction device 1002, the non-recursive filter noise reduction device 1003, and the contour correction device 1006. Is what you do. It should be noted that the motion detection signal and the video signal level are determined by the multiplier 1 shown in FIG.
It can be extracted from the signal input to 105.

The contour correction device 1006 is connected in parallel with the non-recursive filter noise reduction device 1003, performs a contour correction process on the signal output from the frame recursive noise reduction device 1002, and outputs a contour correction signal. It is. The adder 1007 includes the video signal output from the non-recursive filter noise reduction device 1003 and the contour correction device 1
006 is added together with the contour correction signal output from 006 and output as an output video signal 1005.

The motion adaptive frame recursive noise reduction apparatus 1002 detects video motion from the input video signal 1001 and increases the amount of recursion of the delayed image when the motion is small, and decreases the recursive amount when the motion is large. In other words, the control is performed so that the effect of noise reduction is greater as the camera is closer to the stationary state, and the afterimage is not generated as the movement increases. When the video signal level is low, the amount of circulation is large, and when the video signal level is high, the amount of circulation is small. From the above, for example, when arranged in descending order of the traveling amount,

Stationary and small level> stationary and large level> moving and small level> moving and large level: Equation (3), and can be controlled by the control device 1004 in this manner.

The frame recursive noise reduction device 10
For a video signal for which noise reduction was not sufficient in 02, the control device 1004 controls the non-recursive filter noise reduction device 1003 to increase the noise reduction effect. That is, control is performed such that the effect of noise reduction increases as the moving image is closer to the moving image. When the video signal level is low, noise is conspicuous, and control is performed to increase the noise reduction effect. From the above, for example, if the noise reduction effect is arranged in descending order,

Moving and low level> moving and high level> still and low level> moving and high level: Equation (4), and can be controlled by the control device 1004 in this manner. The control means 1004 can be applied not only to the acyclic noise reduction device for adaptively selecting and averaging pixels shown in the conventional example but also to various acyclic filters, and the effect can be obtained. It is something that can be done. The above is the first
An example of control similar to that of the embodiment is shown.

The contour correction device 1006 includes a control unit 100
4, the operation of which is controlled by the non-recursive filter noise reduction device 1003.
An operation is performed to reduce the blurring of the outline portion 05.

In general, when a motion occurs, the amplitude of the contour correction signal decreases. Therefore, when the motion occurs, the gain of the contour correction device is increased to increase the amplitude of the contour correction signal. Also,
Since noise when the video signal level is low is conspicuous, the gain of the contour correction device is reduced to reduce the amplitude of the contour correction signal. From the above, for example, when arranged in the order of increasing the gain of the contour correction signal,

Moving and high level> moving and low level> still and high level> still and low level ... The control means 1004 controls the contour correction signal so as to obtain the following equation (5), and the contour is blurred when moving. , And the noise of the low-level contour correction signal when stationary can be reduced, and the output video signal 1005 can be obtained.

In general, the gradient of the contour waveform at the time of stationary is steep, and the gradient of the contour waveform of the video signal at the time of motion becomes gentle. Therefore, by making the gradient at the time of motion steep, Control can also be performed to reduce motion blur. If the operation of increasing the gradient is performed when the video signal level is high due to only the movement, the contour blur can be reduced and the image can be sharpened. When arranged in descending order of steepness,

Large motion and large level> Large motion and small level> Small motion and large level> Small motion and small level Expression (6) As described above, the control means 1004 can control the contour correction device 1006. it can.

As described above, the noise reduction apparatus according to the second embodiment of the present invention is a frame recursive noise reduction apparatus 1002 that performs noise reduction on an input video signal 1001.
And a control unit 1004 to which a motion detection signal and a video signal level obtained from the frame recursive noise reduction device 1002 are input.
2 is connected in series with a non-recursive filter noise reduction device 1003 controlled by the control means 1004, and a contour correction device 1006 controlled by the control means 1004.
Is connected in parallel with the non-recursive filter noise reduction device 1003, so that afterimages do not occur when there is motion in consideration of the motion of the video and the video signal level, and when the video is still, the level is low. It is possible to mainly reduce noise of a video signal. In addition, there is an effect that the outline of the video is sharpened without emphasizing noise according to the motion and the level. (Third embodiment)

Next, FIG. 3 shows a configuration diagram of a noise reduction device according to the third embodiment.

As shown in FIG. 3, the noise reduction device includes a frame recursive type noise reduction device 1002, a non-recursive type filter noise reduction device 1003, control means 1004 and A
It has a GC device 1008, receives an input video signal 1001, performs noise reduction, and outputs an output video signal 1005.

Frame cyclic noise reduction apparatus 1002
Is for inputting a video signal from the AGC device 1008, performing noise reduction processing based on motion detection information, and outputting the result. Non-recursive filter noise reduction device 1003
Is connected in series with the frame recursive noise reduction apparatus 1002, adaptively selects pixels and averages the signal output from the frame recursive noise reduction apparatus 1002, and performs noise reduction. This is output as 1005.

The control device 1004 receives the motion detection signal from the frame recursive noise reduction device 1002 and
The gain amount is input from the device 1008, and the frame recursive noise reduction device 1002 and the non-recursive filter noise reduction device 1003 are controlled. Note that the motion detection signal can be extracted from the signal input to the multiplier 1105 shown in FIG.

The AGC device 1008 is connected in series with the frame recursive noise reduction device 1002, receives the input video signal 1001, and automatically adjusts the gain so that a predetermined output signal can be maintained even if the input signal strength changes. In order to make it easier to see the dark part of the input video signal 1001 and output it to the frame recursive noise reduction device 1002,
The gain amount of C (automatic gain control) is output to the control means 1004.

The control means 1004 controls the motion adaptive frame recursive noise reduction device 1002 to operate the AGC device 10.
08 to detect the motion of the video signal output from the video signal,
When the movement is small, the amount of circulation of the delayed image is controlled to be large, and when the movement is large, the amount of circulation is controlled to be small, that is, the effect of noise reduction is large as the movement is close to static, and the afterimage does not occur as the movement becomes large. To control. When the gain of the output of the AGC device 1008 is small, that is, when the input video signal 1001 is large, the amount of repetition is small. When the gain of the output of the AGC device 1008 is large, that is, when the input video signal 1001 is small, noise is reduced. To increase the traveling amount. From the above, for example, when arranged in descending order of the traveling amount,

Stationary and large gain> static and small gain> moving and large gain> moving and small gain Equation (7), and thus the control means 1004 can control the frame recursive noise reduction apparatus 1002. .

The frame recursive noise reduction device 10
For video signals for which noise reduction was not sufficient in 02, the control unit 1004 controls the non-recursive filter noise reduction device 1003 to increase the noise reduction effect. That is, control is performed such that the effect of noise reduction increases as the movement increases. Also, the AGC device 100
When the gain of 8 outputs is large, noise is noticeable.
Control is performed to increase the noise reduction effect. From the above,
For example, if the noise reduction effect is arranged in descending order,

Motion and large gain> movement and small gain> stationary and large gain> stationary and small gain Equation (8), and the control means 1004 controls the non-recursive filter noise reduction apparatus 1003 in this way. Thus, an output video signal 1005 can be obtained.

As described above, the noise reduction device according to the third embodiment of the present invention is provided with the AGC device 1008 for making the dark part of the input video signal 1001 easily visible, and is connected to the AGC device 1008 in series. A frame recursive noise reduction device 1002 and a control unit 1004 controlled by an AGC gain amount obtained from the AGC device 1008 and a motion detection signal obtained from the frame recursive noise reduction device 1002 are provided. Acyclic filter noise reduction apparatus 1003
Are connected in series to the frame recursive noise reduction apparatus 1002, so that it is possible to reduce noise using both the motion of the video and the amount of AGC gain. Further, the control means 1004 is applicable not only to the noise reduction device described in the conventional example but also to various non-recursive filters,
An effect can be obtained. (Fourth embodiment)

Next, FIG. 4 shows a configuration diagram of a noise reduction device according to the fourth embodiment.

As shown in FIG. 4, the noise reduction apparatus includes a frame recursive noise reduction apparatus 1002, a non-recursive filter noise reduction apparatus 1003, and a region control means 1009.
The input video signal 1001 is input, noise is reduced, and the output video signal 1005 is output.

Frame Cyclic Noise Reduction Apparatus 1002
Is for performing a noise reduction process on the input video signal 1001 based on the motion detection information and outputting the result. The non-recursive filter noise reduction device 1003 is connected in series with the frame recursive noise reduction device 1002, and adaptively selects and averages pixels output from the frame recursive noise reduction device 1002 by averaging. The noise is reduced and output.

The area control means 1009 receives the motion detection signal output from the frame recursive noise reduction apparatus 1002 and controls the operation of the non-recursive filter noise reduction apparatus 1003. Note that the motion detection signal can be extracted from the signal input to the multiplier 1105 shown in FIG.

Frame cyclic noise reduction apparatus 1002
Reduces the amount of recursion so that the afterimage does not occur as the motion of the video of the input video signal 1001 increases, so that the noise reduction effect decreases. Therefore, the input video signal 10
As the movement of the image of the image No. 01 increases, the area control means 1
009, the acyclic filter noise reduction device 100
By performing control so as to increase the number of peripheral pixels involved in the operation for replacing the center pixel by expanding the region where the filtering process of No. 3 is performed, effective noise reduction is performed even during motion. As the movement becomes smaller and approaches stationary, the area control unit 1009 controls the non-recursive filter noise reduction device 1003 so as to narrow the area where the non-cyclic filter processing is performed. In other words, by reducing the number of peripheral pixels involved in the operation for replacing the central pixel, noise reduction is performed only on correlated pixels, so that noise can be reduced by suppressing blurring of the contour. Thus, an output video signal 1005 can be obtained.

For example, a case where the average value of peripheral pixels is replaced with a central pixel will be described. When the movement is small, the number N1 of peripheral pixels surrounding the central pixel involved in the calculation (pixel value:
XA1, XA2,... XAN1), and N2 pixels (pixel values: YA1, YA2,.
And When the movement increases, the peripheral pixels to be calculated spread further from the center pixel toward the periphery, so that N1 <N2
Becomes The average value of surrounding pixels when motion is small is

(XA1 + XA2 +... XAN1) / N1 (9) The average value when the motion is large is

(YA1 + YA2 +... YAN2) / N2... (10) and can be replaced with the center pixel. Although the above description has been made in two steps, three or more steps can be used for more detailed control. Further, the above is only an example, and the center pixel can be replaced by another calculation method.

As described above, the noise reduction apparatus according to the fourth embodiment of the present invention is a frame recursive noise reduction apparatus 1002 that performs noise reduction on an input video signal 1001.
Area control means 10 controlled by a motion detection signal output from frame recursive noise reduction apparatus 1002
09, and the non-recursive filter noise reduction device 1003 that operates under the control of the region control means 1009 is provided in series with the frame recursive noise reduction device 1002, so that the region of the filter can be determined using the motion detection signal of the video signal. By changing the noise, effective noise reduction can be performed. The area control unit 1009 is applicable not only to the noise reduction device described in the conventional example but also to various non-recursive filters, and an effect can be obtained. (Fifth embodiment)

Next, FIG. 5 shows a configuration diagram of a noise reduction device according to the fifth embodiment.

As shown in FIG. 5, the noise reduction apparatus includes a frame recursive noise reduction apparatus 1002, a non-recursive filter noise reduction apparatus 1003, and a distance control means 1010.
The input video signal 1001 is input, noise is reduced, and the output video signal 1005 is output.

In this embodiment, the distance control means 1010 is used instead of the area control means 1009 in the above embodiment.
10 is provided.

The distance control means 1010 controls the non-recursive filter noise reduction device 1003, and weights and controls the pixels in the filter operation for replacing the center pixel.

Frame cyclic noise reduction apparatus 1002
Reduces the amount of reciprocation so that the afterimage does not occur as the motion of the video of the input video signal 1001 increases, so that the noise reduction effect decreases. Therefore, the distance control means 10
10, the non-recursive filter noise reduction device 1003
The weights of the pixels involved in the filter operation to replace the center pixel are changed according to the distance from the center pixel. In other words, the weight of the pixel closer to the center pixel is increased, and the weight of the pixel farther from the center pixel is reduced.

Further, the weight is changed depending on the pixel position according to the magnitude of the motion, and as the motion increases, the range of the pixel with the higher weight is expanded from the center pixel to the peripheral direction, and the area of the pixel with the lower weight is a static area. The distance is farther from the center pixel than when it is. When the range of pixels having a large weight is expanded, the range of pixels to be calculated having a correlation with the center pixel is expanded, so that effective noise reduction can be performed even in motion.

Further, the range of the pixels having a large weight is reduced to the vicinity of the center pixel as the image approaches the stationary state, and the range of the pixels having a small weight is reduced to the center pixel side as compared with the case of the motion. As a result, pixels far from the pixel close to the central pixel are regarded as having low correlation, and noise can be reduced by focusing on pixels near the central pixel. Thus, an output video signal 1005 can be obtained.

For example, a case where the weighted average value is replaced with the center pixel will be described as an example. Let the weighting value of a pixel near the center pixel be A, and let the weighting value of a pixel far from the center pixel be B
(A> B).

The number of pixels corresponding to the weighting value A is a small number N1 (pixel values: XA1, XA2,..., XAN1) surrounding the central pixel when the motion is small, and N2 (pixel values) when the motion is large. : YA1, YA2, ... YAN
2).

When the motion is small, M1 pixels (pixel values: XB1, XB2,..., XBM) correspond to the weighting value B.
1), and when the movement increases, M2 pixels (pixel value:
YB1, YB2,... YBM2).

Therefore, spatially, M1 and M2 pixels are allocated around the N1 and N2 pixels, respectively. When the movement increases, the area of the weighting value A spreads from the center pixel to the periphery. , N1 <N2.
When the movement is small, the weighted average is

{A (XA1 + XA2 +... XAN1) + B (XB1 + XB2 +... XBM1)} / (A × N1 + B × M1) Formula (11)

{A (YA1 + YA2 +... YAN2) + B (YB1 + YB2 +... YBM2)} / (A × N2 + B × M2) Formula (12), which can be replaced with the center pixel. The above is
Although the weighting is performed in two stages, it may be performed in three or more stages in order to perform more detailed weighting. In addition to the above method, the center pixel can be replaced using weighting.

As described above, the noise reduction apparatus according to the fifth embodiment of the present invention is a frame recursive noise reduction apparatus 1002 that performs noise reduction on an input video signal 1001.
And distance control means 10 controlled by a motion detection signal output from frame recursive noise reduction apparatus 1002
10, and a non-recursive filter noise reduction device 1003 operating under the control of the distance control means 1010 is provided in series with the frame recursive noise reduction device 1002, so that it is attached to peripheral pixels using the motion detection signal of the video signal. By adjusting the weight, effective noise reduction can be achieved. The distance control means 1010 is applicable not only to the noise reduction device described in the conventional example but also to various non-recursive filters, and an effect can be obtained. (Sixth embodiment)

Next, FIG. 6 shows a flowchart of the noise reduction method according to the sixth embodiment.

As shown in the flowchart of FIG. 6, the noise reduction method of this embodiment captures an image in step 1 (S1) and calculates the difference between the image captured in step 2 and the image captured one frame before. Then, the motion is detected (S2), and the level of the video captured in step 3 is detected (S3).

Subsequently, in step 4, the result of the motion detection in S2 and
The cyclic amount of the frame is determined based on the detection result of the video signal level of S3 so as to perform the control of the equation (1) (S4). The control parameters of the recursive filter are determined so as to control the equation (2) (S5).

Then, cyclic noise reduction is performed based on the cyclic amount determined in step S4 in step 6 (S6), and in step 7,
Noise is reduced by the non-recursive filter processing using the control parameters determined in S5, and the result is set as the final output (S7).

As described above, in the noise reduction method according to the sixth embodiment of the present invention, the amount of control and the control parameters are determined based on the motion detection (S2) and the video signal level detection (S3).
(S4, S5), recursive noise reduction (S6) and acyclic filter noise reduction (S7). In the stationary state, noise reduction can be performed with emphasis on a signal having a small level. (Seventh embodiment)

Next, FIG. 7 shows a flowchart of the noise reduction method according to the seventh embodiment.

As shown in the flowchart of FIG. 7, in the noise reduction method according to the present embodiment, an image is captured in step 1 (S1), and a difference between the image captured in step 2 and the image captured one frame before is calculated. Then, the motion is detected (S2), and the level of the video captured in step 3 is detected (S3).

Subsequently, in step 4, the result of the motion detection in S2 is
The cyclic amount of the frame is determined based on the detection result of the video signal level of S3 so as to perform the control of Expression (3) (S4). The control parameters of the recursive filter are determined so as to control the equation (4) (S5).

Further, in step 8, the motion detection result of S2 and
Based on the detection result of the video signal level in S3, control parameters for steepening the gain amount or gradient of the contour correction are determined so as to perform the control of Expression (5) or Expression (6) (S8).
In step 6, cyclic noise reduction is performed based on the cyclic amount determined in S4 (S6).

Next, in step 9, an outline correction signal is generated based on the control parameters determined in S8 (S9), and in step 7, non-recursive filter noise reduction is performed using the control parameters determined in S5 (S7). ), Both the contour correction signal obtained in S9 and the noise-reduced signal obtained in S7 are added as a final output (S10).

As described above, the noise reduction method according to the seventh embodiment of the present invention uses the amount of recursion and the control parameters of the nonrecursive filter based on the motion detection (S2) and the video signal level detection (S3). And control parameters for contour correction (S4, S4
5, S8), recursive noise reduction (S6), non-recursive filter noise reduction (S7) and contour correction (S9), so taking into account the motion of the video signal and the video signal level, Noise reduction is performed by a non-recursive filter with no afterimages. When the image is stationary, noise is reduced by a recursive filter for low-level signals, and contour blur is reduced even if motion blur is reduced and noise is reduced. be able to. (Eighth Embodiment)

Next, FIG. 8 shows a flowchart of the noise reduction method according to the eighth embodiment.

As shown in the flowchart of FIG. 8, in the noise reduction method of this embodiment, an image is captured in step 1 (S1), and an AGC (Step 1) is performed so that a dark portion of the image captured in step S1 is easily seen. A gain amount for performing automatic gain control) is calculated (S11), and based on the gain amount obtained in S11 in step 12, the video signal captured in S1 is converted to A
Amplified by GC (S12), the difference between the video captured in S1 in step 2 and the video captured one frame before in S1 or the video amplified by AGC in S12 and one frame before amplified in S12 Motion detection is performed by calculating the difference from the video (S2).

Subsequently, in step 4, the motion detection result of S2 and
The cyclic amount of the frame is determined based on the AGC gain amount obtained in S11 so as to control the equation (7) (S4), and step 5 is performed.
Then, the control parameters of the non-recursive filter are determined based on the motion detection result in S2 and the AGC gain calculated in S11 so as to be controlled by the equation (8) (S5).

Then, cyclic noise reduction is performed in step 6 based on the cyclic amount determined in S4 (S6), and in step 7,
Non-recursive filter noise reduction is performed using the control parameters determined in S5 to obtain the final output (S7).

As described above, in the noise reduction method according to the eighth embodiment of the present invention, the motion detection (S2) and the AGC gain calculation are performed.
Based on (S11), determine the amount of circulation and control parameters
(S4, S5), to perform the cyclic noise reduction (S6) and the non-recursive filter noise reduction (S7), so as to improve the phenomenon that the AGC increases the low level video signal amplitude and deteriorates the S / N. In addition, by mainly performing cyclic noise reduction in stationary areas and mainly performing noise reduction using a non-recursive filter in the case of motion, low-level signals amplified by AGC are mainly reduced in noise. Can be. (Ninth embodiment)

Next, a flowchart of the noise reduction method according to the ninth embodiment is shown in FIG.

As shown in the flowchart of FIG. 9, the noise reduction method of this embodiment captures an image in step 1 (S1) and calculates the difference between the image captured in step 2 and the image captured one frame before. Then, motion detection is performed (S2).

Next, in step 4, the amount of recursion of the frame is determined from the motion detection result of S2 (S4). In step 13, the region control parameters of the non-recursive filter are determined from the motion detection result of S2. The area is determined to be a small area only around the filter center pixel, and the control is such that the filter area is controlled to expand from a small area having only the filter center pixel as the movement increases (S13).

Then, cyclic noise reduction is performed based on the cyclic amount determined in S4 in step 6 (S6), and in step 7,
Non-recursive filter noise reduction is performed using the control parameters determined in S13, and the result is set as the final output (S7).

As described above, in the noise reduction method according to the ninth embodiment of the present invention, the cyclic amount and the control parameters are determined based on the motion detection (S2) (S4, S13), and the cyclic noise reduction is performed. (S6) and acyclic filter noise reduction (S7)
In the case of motion of the video signal, the range of correlated pixels is expanded in the non-recursive filter processing to perform filter processing to perform noise reduction with no afterimage, and in the case of stillness, to perform cyclic noise reduction. In these two types of noise reduction, signals with low levels can be mainly reduced. (Tenth embodiment)

Next, a flowchart of the noise reduction method according to the tenth embodiment is shown in FIG.

As shown in the flowchart of FIG. 10, in the noise reduction method of this embodiment, a video is captured in step 1 (S1), and a difference between the video captured in step 2 and the video captured one frame before is calculated. Then, motion detection is performed (S2).

Subsequently, in step 4, the amount of recursion of the frame is determined from the motion detection result of S2 (S4). In step 14, the weighting control parameter of the acyclic filter is determined from the motion detection result of S2. The area where the weight of is large is a small area only around the filter center pixel,
The weight decreases as the distance from the filter center pixel increases, and as the motion increases, the area with the larger weight is enlarged from the small area of only the filter center pixel, and the area with the smaller weight is further increased from the filter center pixel. It is determined that the control will be made far (S14).

Then, cyclic noise reduction is performed based on the cyclic amount determined in step S4 in step 6 (S6), and in step 7,
Non-recursive filter noise reduction is performed using the control parameters determined in S14 to obtain the final output (S7).

As described above, in the noise reduction method according to the tenth embodiment of the present invention, the cyclic amount and the control parameters are determined based on the motion detection (S2) (S4, S14), and the cyclic noise reduction is performed. (S6) and non-recursive filter noise reduction (S7), so in the case of video signal motion, the area of correlated pixels in the non-recursive filter processing, that is, the area of the filter processing weight is expanded from the filter center pixel. By doing so, it is possible to perform noise reduction with the effect of eliminating afterimages using a non-recursive filter when there is motion, and to effectively reduce noise at low levels using a cyclic filter when stationary. it can.

[0128]

As described above, the present invention provides a cyclic noise reduction apparatus for reducing the noise of an input video signal and outputting the same, and a video signal input from the cyclic noise reduction apparatus to reduce noise. Since the control means controls the output of the non-recursive filter noise reduction device based on the motion detection signal and the video signal level, there is no afterimage where there is motion in consideration of the video motion and the video signal level, A noise reduction device that has an excellent effect of being able to mainly reduce noise of a signal having a small level when stationary.
It is possible to provide a medium on which a noise reduction method and a noise reduction program are recorded.

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of a noise reduction device according to the present invention.

FIG. 2 is a block diagram illustrating a configuration of a noise reduction device according to an embodiment.

FIG. 3 is a block diagram illustrating a configuration of a noise reduction device according to an embodiment.

FIG. 4 is a block diagram illustrating a configuration of a noise reduction device according to an embodiment.

FIG. 5 is a block diagram illustrating a configuration of a noise reduction device according to an embodiment.

FIG. 6 is a flowchart illustrating a noise reduction method according to an embodiment.

FIG. 7 is a flowchart illustrating a noise reduction method according to an embodiment.

FIG. 8 is a flowchart illustrating a noise reduction method according to an embodiment.

FIG. 9 is a flowchart illustrating a noise reduction method according to an embodiment.

FIG. 10 is a flowchart illustrating a noise reduction method according to an embodiment.

FIG. 11 is a block diagram illustrating an example of a conventional cyclic noise reduction device.

FIG. 12 is a block diagram illustrating an example of a conventional acyclic noise reduction device.

[Explanation of symbols]

Reference Signs List 1001 input video signal 1002 frame recursive noise reduction device 1003 non-recursive filter noise reduction device 1004 control means 1005 output video signal 1006 contour correction device 1007 adder 1008 AGC device 1009 area control means 1010 distance control means 1101 input video signal 1102 2 Dimension LPF (Low Pass Filter) 1103 Adder 1104 Motion Judgment Circuit 1105, 1106 Multiplier 1107 Adder 1108 Frame Memory 1109 Output Video Signal 1201 Input Video Signal 1202 Pixel Capture Circuit 1203, 1204, 1205, 1206 Difference Calculation Circuit 1207 Threshold 1208, 1209, 1210, 1211 Correlated pixel determination circuit 1212 Correlated pixel addition circuit 1213 Pixel number counting circuit 1214 Divider 1215 Output image Signal

 ────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Misa Kasahara 4-3-1 Tsunashimahigashi, Kohoku-ku, Yokohama-shi, Kanagawa F-term (reference) in Matsushita Communication Industrial Co., Ltd. 5C021 PA66 PA79 RA01 RA12 RB01 RB03 RB06 RB07 RB08 XB03 YA01

Claims (15)

[Claims] 1. A recursive type that receives a video signal, adjusts a cyclic amount of a delayed image based on an image acquired from the video signal and a delayed image, reduces noise of the input video signal, and outputs the reduced signal. A noise reduction device, a non-recursive filter noise reduction device that inputs a video signal from the recursive noise reduction device, performs a filtering process using peripheral pixels, and reduces and outputs noise of the input video signal; Control means for controlling the recursive noise reduction device and the non-recursive filter noise reduction device, wherein the recursive noise reduction device detects a motion of a video based on the image and the delayed image to perform motion. Motion determination means for outputting a detection signal, video signal level detection means for detecting and outputting a video signal level of the input video signal, the motion detection signal and the video signal A revolving amount adjusting means for adjusting the revolving amount of the delayed image based on the level, and reducing the noise of the input video signal by revolving the delayed image according to the revolving amount adjusted by the revolving amount adjusting means. And a video signal output unit for outputting the recursive noise reduction device and the non-recursive filter noise reduction device based on the motion detection signal and the video signal level. A noise reduction device characterized by the above-mentioned. 2. A cyclic type which receives a video signal, adjusts the amount of circulation of a delayed image based on an image acquired from the video signal and a delayed image, and reduces noise of the input video signal to output the same. A noise reduction device, a non-recursive filter noise reduction device that inputs a video signal from the recursive noise reduction device, performs a filtering process using surrounding pixels, and reduces and outputs noise of the input video signal. A video signal before noise reduction is input, the contour correction of the image is performed, a contour correction device that reduces and outputs noise of the input video signal, and a video signal output by the non-recursive filter noise reduction device, Video signal output means for generating and outputting an output video signal from the video signal output by the contour correction device; the cyclic noise reduction device; and the non-cyclic filter noise reduction device. And a control unit for controlling the contour correction device, wherein the cyclic noise reduction device detects a motion of a video based on the image and the delayed image and outputs a motion detection signal, and a motion determination unit. Video signal level detecting means for detecting and outputting a video signal level of the input video signal; and cyclic amount adjusting means for adjusting the cyclic amount of the delayed image based on the motion detection signal and the video signal level. And a video signal output unit that circulates the delayed image according to the circulating amount adjusted by the circulating amount adjusting unit, reduces noise of the input video signal, and outputs the reduced image signal. Controlling the recursive noise reduction device, the non-recursive filter noise reduction device, and the contour correction device based on a motion detection signal and the video signal level. Noise reduction apparatus according to symptoms. 3. An automatic gain control device for inputting a video signal and automatically changing a gain while maintaining a predetermined output signal strength, and an image obtained by inputting a video signal from the automatic gain control device and acquiring the video signal. A cyclic noise reduction device that adjusts a cyclic amount of a delayed image based on the delayed image and reduces and outputs noise of an input video signal; and a video signal from the cyclic noise reduction device, A non-recursive filter noise reduction device that performs filter processing using peripheral pixels to reduce and output noise of an input video signal, and control for controlling the recursive noise reduction device and the non-recursive filter noise reduction device Means, wherein the automatic gain control device has gain amount output means for outputting a gain amount obtained by changing the gain of the input video signal, and the recursive noise reduction device comprises: A motion determining unit that detects a motion of a video based on an image and the delayed image and outputs a motion detection signal, and a cyclic amount adjusting unit that adjusts a cyclic amount of the delayed image based on the motion detection signal. Video signal output means for circulating the delayed image in accordance with the circulating amount adjusted by the circulating amount adjusting means, and reducing and outputting noise of the input video signal, wherein the control means comprises: A noise reduction apparatus comprising: controlling the recursive noise reduction apparatus and the non-recursive filter noise reduction apparatus based on an amount and the motion detection signal. 4. A cyclic type which receives a video signal, adjusts a cyclic amount of a delayed image based on an image obtained from the video signal and a delayed image, reduces noise of the input video signal, and outputs the resultant signal. A noise reduction device, a non-recursive filter noise reduction device that inputs a video signal from the recursive noise reduction device, performs a filtering process using peripheral pixels, and reduces and outputs noise of the input video signal; Control means for controlling a non-recursive filter noise reduction device, wherein the recursive noise reduction device detects motion of a video based on the image and the delayed image and outputs a motion detection signal. Means, a cyclic amount adjusting means for adjusting the cyclic amount of the delayed image based on the motion detection signal, and a cyclic amount adjusted by the cyclic amount adjusting means. Video signal output means for circulating an image to reduce the noise of the input video signal and outputting the reduced signal, wherein the control means performs a filter performed by the non-recursive filter noise reduction apparatus based on the motion detection signal. A noise reduction device for controlling an adaptive range of processing. 5. A cyclic type which receives a video signal, adjusts a cyclic amount of a delayed image based on an image obtained from the video signal and a delayed image, reduces noise of the input video signal, and outputs the reduced signal. A noise reduction device, a non-recursive filter noise reduction device that inputs a video signal from the recursive noise reduction device, performs a filtering process using peripheral pixels, and reduces and outputs noise of the input video signal; Control means for controlling a non-recursive filter noise reduction device, wherein the recursive noise reduction device detects motion of a video based on the image and the delayed image and outputs a motion detection signal. Means, a cyclic amount adjusting means for adjusting the cyclic amount of the delayed image based on the motion detection signal, and a cyclic amount adjusted by the cyclic amount adjusting means. Video signal output means for circulating an image to reduce the noise of the input video signal and outputting the reduced signal, wherein the control means performs a filter performed by the non-recursive filter noise reduction apparatus based on the motion detection signal. A noise reduction device characterized in that adaptive pixels for processing are weighted and processed. 6. An image signal inputting step of inputting an image signal, an image delaying step of holding and delaying and outputting an image, acquiring an image from the input image signal, and acquiring the image, A motion detection step of detecting a motion of a video based on the signal level detection step of detecting a signal level of the input video signal, and the delay based on the motion of the video and the video signal level. A cyclic amount determining step of determining a cyclic amount of an image; a filter control parameter determining step of determining a filter control parameter for performing a filtering process on a video signal based on the motion of the video and the video signal level; Cyclic noise that controls a cyclic amount of the delayed image based on the cyclic amount and reduces noise of the input video signal and outputs the noise. A non-recursive filter noise reduction that outputs the video signal output in the recursive noise reduction step by selecting peripheral pixels based on the filter control parameters and performs the filter processing to reduce the noise. And a noise reduction method. 7. A video signal inputting step of inputting a video signal, an image delaying step of holding and delaying and outputting an image, obtaining an image from the input video signal, and obtaining the image, A motion detection step of detecting a motion of a video based on the signal level detection step of detecting a signal level of the input video signal, and the delay based on the motion of the video and the video signal level. A cyclic amount determining step of determining a cyclic amount of an image; a filter control parameter determining step of determining a filter control parameter for performing a filtering process on a video signal based on the motion of the video and the video signal level; An outline correction parameter for determining an outline correction parameter for performing an outline correction process on a video signal based on a motion and the video signal level. A cyclic noise reduction step of controlling a cyclic amount of the delayed image based on the determined cyclic amount, and reducing and outputting noise of the input video signal; A non-recursive filter noise reduction step of selecting peripheral pixels based on the filter control parameters, performing a filtering process on the video signal output in the step, and reducing and outputting noise; and the recursive noise reduction step. From the video signal output in the above, performing the contour correction processing based on the contour correction parameter, a contour correction signal generating step of generating a contour correction signal, and the video signal output in the non-recursive filter noise reduction step And an addition step of adding the contour correction signal. 8. A video signal inputting step of inputting a video signal; a gain calculating step of calculating a gain for maintaining the input video signal at a predetermined output signal strength; An automatic gain control step of amplifying an input video signal; an image delay step of holding and delaying and outputting an image; obtaining an image from the amplified video signal,
A motion detection step of detecting a motion of a video based on the delayed image; and a cyclic amount determining step of determining a cyclic amount of the delayed image based on the gain amount and the motion of the video. A filter control parameter determining step of determining a filter control parameter for performing a filtering process on a video signal based on the gain amount and the motion of the video; and a cyclic operation of the delayed image based on the determined cyclic amount. A cyclic noise reduction step of controlling the amount and reducing the noise of the amplified video signal, and outputting the video signal output in the cyclic noise reduction step to peripheral pixels based on the filter control parameter. A non-recursive filter noise reduction step of selecting and performing the filter processing and reducing and outputting the noise. Noise reduction method. 9. A video signal inputting step of inputting a video signal, an image delaying step of holding and delaying and outputting an image, acquiring an image from the input video signal, and obtaining the image, A motion detection step of detecting a motion of a video, based on the motion of the video, a cyclic amount determining step of determining a cyclic amount of the delayed image based on the motion of the video, and a video signal based on the motion of the video. A filter region control parameter determining step of determining a filter region for performing a filtering process, controlling a cyclic amount of the delayed image based on the determined cyclic amount, and reducing and outputting noise of the input video signal. Recursive noise reduction step, and the video signal output in the recursive noise reduction step is used to determine peripheral pixels based on the filter area control parameter. A non-recursive filter noise reduction step of selecting and performing the filter processing to reduce and output noise. 10. A video signal inputting step of inputting a video signal, an image delaying step of holding and delaying and outputting an image, obtaining an image from the input video signal, and obtaining the image, the delayed image, A motion detection step of detecting a motion of a video, based on the motion of the video, a cyclic amount determining step of determining a cyclic amount of the delayed image based on the motion of the video, and a video signal based on the motion of the video. A filter weighting control parameter determining step of determining a weight for the pixel to be subjected to the filtering process, and controlling a cyclic amount of the delayed image based on the determined cyclic amount to reduce noise of the input video signal. A cyclic noise reduction step to be output; and a video signal output in the cyclic noise reduction step, based on the filter weight control parameters. A non-recursive filter noise reduction step of performing weighting on peripheral pixels to perform the filtering process and reducing and outputting noise. 11. A motion detecting step of obtaining an image from a video signal and detecting a motion of the video based on the image and the delayed image; and a signal level detecting step of detecting a signal level of the video signal. Determining a cyclic amount of the delayed image based on the motion of the video and the video signal level; and filtering the video signal based on the motion of the video and the video signal level. A filter control parameter determining step of determining a filter control parameter to be processed; and a cyclic noise reduction step of controlling a cyclic amount of the delayed image based on the determined cyclic amount to reduce noise of the video signal. And selecting a peripheral pixel based on the filter control parameter from the video signal whose noise has been reduced in the cyclic noise reduction step. Wherein to filter, medium recording noise reduction program, characterized in that a non-recursive filter noise reducing step of reducing noise Te. 12. A motion detecting step for obtaining an image from a video signal and detecting a motion of the video based on the image and the delayed image; a signal level detecting step for detecting a signal level of the video signal. Determining a cyclic amount of the delayed image based on the motion of the video and the video signal level; and filtering the video signal based on the motion of the video and the video signal level. A filter control parameter determining step of determining a filter control parameter for performing processing; and a contour correction parameter determining step of determining a contour correction parameter for performing contour correction processing on a video signal based on the motion of the video and the video signal level. Controlling a cyclic amount of the delayed image based on the determined cyclic amount to reduce noise of the video signal; A recursive noise reduction step, and a non-recursive filter noise for reducing the noise by selecting peripheral pixels based on the filter control parameters and performing the filtering process on the video signal reduced in the recursive noise reduction step. A contour correction signal generating step of performing a contour correction process based on the contour correction parameter from the video signal noise reduced in the cyclic noise reduction step to generate a contour correction signal; A medium recording a noise reduction program, comprising: an addition step of adding a video signal whose noise has been reduced in a filter noise reduction step and the contour correction signal. 13. A gain amount calculating step of calculating a gain amount for maintaining a video signal at a predetermined output signal strength; an automatic gain control step of amplifying the video signal based on the gain amount; Acquire an image from the video signal and
Based on the delayed image, a motion detection step of detecting the motion of the video, Based on the gain amount and the motion of the video, a cyclic amount determining step of determining the cyclic amount of the delayed image, A filter control parameter determining step of determining a filter control parameter for performing a filtering process on a video signal based on the gain amount and the motion of the video; and a cyclic amount of the delayed image based on the determined cyclic amount. A cyclic noise reduction step of reducing the noise of the amplified video signal, and a video signal whose noise has been reduced in the cyclic noise reduction step selects peripheral pixels based on the filter control parameters. A noise reduction step of performing a filtering process to reduce noise. Medium on which program is recorded. 14. A motion detecting step of acquiring an image from a video signal and detecting a motion of the video based on the image and the delayed image; and detecting the motion of the delayed image based on the motion of the video. A cyclic amount determining step of determining a cyclic amount; a filter region control parameter determining step of determining a filter region for performing a filtering process on a video signal based on the motion of the video; and the delay based on the determined cyclic amount. A recursive noise reduction step of controlling the amount of repetition of the obtained image and reducing noise of the video signal, and converting the video signal noise reduced in the recursive noise reduction step into peripheral pixels based on the filter area control parameter And a non-recursive filter noise reduction step of performing the filter processing to reduce noise. The recorded medium's reduction program. 15. A motion detecting step of acquiring an image from a video signal and detecting a motion of the video based on the image and the delayed image; and detecting the motion of the delayed image based on the motion of the video. A cyclic amount determining step of determining a cyclic amount; a filter weight control parameter determining step of determining a weight for a pixel for performing a filtering process on a video signal based on the motion of the video; and A cyclic noise reduction step of controlling a cyclic amount of the delayed image to reduce noise of the video signal; and a video signal reduced in the cyclic noise reduction step based on the filter weighting control parameter. Performing the filter processing by weighting the pixels,
And a non-recursive filter noise reduction step for reducing noise.